Intra-Skip Mode Prediction Codeword Allocation in 3D-HEVC

The patent implements dynamic codeword allocation by determining the occurrence probability of each intra-prediction method based on neighboring block statistics, then assigning shorter codewords to high-probability methods and longer codewords to low-probability methods. This dynamic adaptation to changing video content characteristics resolves the contradiction by maintaining low computational complexity while significantly improving coding efficiency through probability-based optimization.

The patent changes the parameter allocation strategy from fixed to adaptive by calculating occurrence probabilities of intra-prediction methods and dynamically assigning index values based on these probabilities. High-probability methods receive lower index values (shorter codewords) while low-probability methods receive higher index values (longer codewords), thereby reducing the average bitstream length and improving compression efficiency while maintaining decoding simplicity through standardized probability-based indexing.

The patent applies partial adaptation by focusing probability calculations only on the most frequently used intra-prediction methods (e.g., vertical, horizontal, diagonal) rather than all possible prediction methods. This selective approach achieves significant coding efficiency improvements for the dominant prediction patterns while limiting the computational overhead, thereby resolving the contradiction between coding efficiency and device complexity through targeted optimization.